Seaglider Pilot's Guide

Seaglider Pilot's GuideSCHOOL OF OCEANOGRAPHY

andAPPLIED PHYSICS LABORATORY

UNIVERSITY OF WASHINGTON

Table of Contents

Seaglider Pilot's Guide 3/1/11 2:06 PM

1 of 18

Chapter 1 — Introduction 1.1 Purpose1.2 Conventions

Chapter 2 — History

Chapter 3 — Principles of Operation 3.1 Density3.2 Forces3.2.1 Static Forces3.2.1.1 Gravity3.2.1.2 Buoyancy3.2.2 Dynamic Forces3.2.2.1 Lift3.2.2.2 Drag3.2.2.3 Hydrodynamic Model3.2.3 Environmental3.2.3.1 Stratification3.2.3.2 Currents3.3 Control of Static Forces3.3.1 Pitch3.3.2 Roll3.3.3 Buoyancy3.4 Features of Control3.4.1 Canonical Dive3.4.2 Control Design3.4.3 Run Phases3.4.3.1 Launch3.4.3.2 Surface3.4.3.3 Dive3.4.3.4 Apogee3.4.3.5 Climb3.4.3.6 Recovery

Chapter 4 — Command and control4.1 Seaglider4.1.1 Serial Communications4.1.2 Power on/off4.1.3 TOM8, PicoDOS, Seaglider Operating Program4.1.4 Menus4.1.5 Extended PicoDOS4.2.1 Function4.2.2 Configuration4.2.3 Files

Chapter 5 — Mission Execution5.1 Planning5.1.1 Environment5.1.1.1 Stratification5.1.1.2 Currents5.1.1.3 Bathymetry5.1.2 Science Requirements5.1.2.1 Track5.1.2.2 Sampling5.1.3 Endurance5.2 Preparation5.2.1 Refurbishment5.2.2 Calibrations5.2.3 Bathymetry Maps5.2.4 Ballasting5.2.5 Harware Checkout and Self-Test5.2.6 Deck Dives5.2.7 Transport5.3 Launch5.4 Test and Trim Dives5.4.1 Diveplot5.4.2 Ballasting5.4.3 Progression of Dives5.5 Monitoring5.5.1 Seaglider Performance5.5.2 Data Completeness and Quality5.5.3 Troubleshooting5.5.4 Resources5.6 Recovery

Chapter 6 — Special Topics6.1 Scenario Run6.1.1 Setup6.1.2 Run Phases6.1.2.1 Launch6.1.2.2 Scene6.1.2.3 Transition6.1.2.4 Recovery6.1.3 Data Access6.2 Ballasting

Chapter 7 — Nomenclature7.1 Terminology7.2 Conventions

Appendix: Table of Conversion Factors

Chapter 1Introduction

1.1 Purpose

This document is an introduction to the use and operation of the University of Washington Seaglider. It provides necessary background andoperating information for new pilots who will be responsible for the operation of Seagliders. This guide is meant to be used closely with thereference manuals. It is not a hardware maintenance or troubleshooting manual, nor does it deal with details of the software that controls theSeaglider's operation.

1.2 Conventions

Parameters that are used to control the operations of Seaglider are given in uppercase bold font, and have leading $ signs ($T_DIVE). Filenames that are used in Seaglider command, control, or operations are given in lowercase bold font (cmdfile). Computer commands given ata prompt, especially when directly connected to the Seaglider are given in italic font (go 2bcf8).


2 of 18

Chapter 2History

The history of buoyancy-driven oceanographic instruments begins with Archimedes (287 BCE - 212 BCE). Archimedes's life began andended in Syracuse, Sicily, but he was educated and spent part of his life in Alexandria, Egypt. He is generally regarded as one of the threegreatest mathematicians of all time (Newton and Gauss complete the triumvirate) and is considered the father of hydrostatics, staticmechanics, and integral calculus. Archimedes's Principle, the most well known of his hydrostatic results, is the basis for all buoyancy-drivenvehicles. It states that the buoyant (upward) force on a submerged object is equal to the weight of the fluid that is displaced by the object.This fact is used in the variable mass, fixed volume (ballast) control systems of modern submarines and submersibles, and in the fixed mass,variable volume control systems of small profiling oceanographic instruments.

The use of buoyancy control in oceanographic instruments dates from the mid-1950's. By 1955, Henry Stommel of the Woods HoleOceanographic Institution and John Swallow in the United Kingdom had ideas for neutrally buoyant floats whose positions could be trackedacoustically. Swallow was the first to build such a device, which contained a free-running 10kHz acoustic source and was tracked from asurface ship. By the 1970's, transponding versions running at 3-4kHz had extended shipboard detection ranges to 50 km, and a 200Hzversion used the Sound Fixing and Ranging (SOFAR) sound channel (Stommel's original idea) to remove the requirement for ship-basedtracking. By the 1980's, Tom Rossby at URI had developed the inverse of the SOFAR float (called RAFOS, SOFAR spelled backwards) thatrelied on moored sound sources and an acoustic receiver on the float. By adding a compressee (an object whose compressibility isapproximately the same as that of seawater), these floats could also be ballasted to follow a particular density surface, rather than a pressuresurface. About the same time, John Dahlen's group at Charles Stark Draper Laboratory developed a moored profiler that used a variablebuoyancy device to propel itself up and down along the mooring wire, measuring temperature, conductivity and currents.

In the 1990's, Russ Davis and his group at Scripps Institution of Oceanography added a variable-buoyancy device to a neutrally buoyant floatto create profiling floats. These floats (called Autonomous Lagrangian Current Explorers, or ALACE) had the ability to inflate an externalbladder, thereby changing their displaced volume, but not their mass. The resulting buoyancy force allowed the float to make profilemeasurements from its neutrally-buoyant depth to the surface. At the surface, position and profile data were transmitted via the ServiceARGOS satellite system. By the year 2000, hundreds of this type of float were deployed worldwide, both of the Scripps design and a designfrom Webb Research Corporation of Falmouth, Massachusetts.

Gliders share a common heritage: Henry Stommel's vision, published in 1989 in Oceanography [Stommel, 1989]. Stommel imagined a fleet ofvehicles that "...migrate vertically through the ocean by changing ballast, and they can be steered horizontally by gliding on wings. Duringbrief moments at the surface, they transmit their accumulated data and receive instructions. Their speed is about 0.5 knot." A prototypegliding vehicle was fielded as early as 1991 by Webb Research Corporation (WRC). This vehicle demonstrated the basic configurationsubsequently used by all three gliders. A few years later, the ONR-sponsored Autonomous Ocean Sensing Network (AOSN) program, led byTom Curtin, sponsored three groups to develop autonomous underwater gliders: Webb Research Corporation, whose glider is called Slocum,a team of Scripps Institution of Oceanography (Russ Davis) and Woods Hole Oceanographic Institution (Breck Owens) who developed Spray,and the University of Washington (Charlie Eriksen) who developed Seaglider. All groups worked with similar design goals: small enough to behandled by two people, relatively low acquisition and operation costs, horizontal speeds of around 30 cm/s, endurance of up to a year, GPSpositioning and two-way data telemetry at the surface, and basic sensor payloads, including a CTD. By the year 2000, all groups hadoperational models that addressed these design goals. This document addresses the details of operation of the University of WashingtonSeaglider, developed jointly between the UW School of Oceanography and the Applied Physics Laboratory.

Chapter 3Principles of Operation

3.1 Density

Density is defined as mass per unit volume:

ρ = M/V, or dimensionally, ρ = M/L3.

Oceanographers routinely switch between SI (mks) and cgs units when referring to seawater densities. Densities are referred to in g/cm3

(with typical 1000m ocean values of 1.0275 ) or kg/m3(with typical 1000m ocean values of 1027.5kg/m3). To further complicate matters,oceanographers have a shorthand notation for density, called σ. σ is defined as follows.

σ = (ρ - 1000)kg/m3 .

So the typical 1000m density is 27.5 in σ units. In particular, we will use the unit σT, which means that the density of the seawater parcel inquestion is computed using the in situ (observed) temperature of the parcel. Densities discussed in glider operations will typically be given incgs units (g/cm3), and we will use the shorthand cc (cubic centimeters) for cm3.

3.2 Forces

3.2.1 Static Forces

Seaglider's flight is controlled by systems that change pitch, roll, and buoyancy. It is designed to operate within a few hundred cc's of neutralbuoyancy over a seawater density range of about 10 σT units. Pitch is controlled to allow upward pitches for climbing, and downward pitchesfor diving and exposing the antenna at the surface. Roll is controlled to effect bank angles, which in turn, cause turns. Pitch and roll arecontrolled by altering the center-of -mass of the vehicle. Buoyancy is controlled by changing the displaced volume of the Seaglider.


3 of 18

3.2.1.1 GravitySeaglider moves a mass inside its pressure hull to change its pitch and roll attitude. The high-voltage (24V) battery pack is axiallyasymmetric. In addition, an 1100g brass weight is mounted on the bottom face of the roughly 6500g pack. This battery and weight is mountedin an assembly called the mass shifter. The battery pack is translated forward and aft to effect changes in vehicle pitch. It is rotated within thepressure hull, and the axial asymmetry effects vehicle roll when the 24V battery pack is rolled from side to side.

Seaglider achieves static trim by addition of ballast (lead weight) inside the fairings. The position and amount of lead is determined by missionand trim requirements. Payload, either additional batteries or sensors, can also change the amount of ballast.

3.2.1.2 BuoyancyBuoyancy is the unbalanced (positive) upward force on a submerged object arising from the vertical pressure gradient. It was Archimedeswho, as mentioned above, stated that the buoyant (upward) force on a submerged object is equal to the weight of the fluid that is displacedby the object.

Submerged objects can alter their buoyancy by changing their density, either by changing their mass or volume. Submarines typically altertheir buoyancy by changing their mass while maintaining their volume. This is done by flooding or blowing seawater from fixed volume ballasttanks.

Gliders, along with ALACE/APEX floats, SOLO floats, the Charles Stark Draper Laboratory Profiling Current Meter (PCM) and others, arefixed-mass, variable-volume devices. They change their buoyancy by changing their displaced volume while keeping their total mass fixed.Typically, this is done by moving hydraulic oil from a reservoir inside a pressure hull to inflate or deflate a rubber bladder external to thepressure hull. The mass of a Seaglider is determined by accurately weighing the full-up instrument in a completely dry condition.

Example 3.1: Suppose a Seaglider has a mass of M = 52,223g. In seawater of density 1.0275g/cm3, what must the Seaglider's volume be inorder for it to be neutrally buoyant at that density?

Answer: From section 3.1.1, V = M/ρ, hence V = 52,223g/1.027g/cm3 = 50,825.3cm3.

Example 3.2: How much must Seaglider's volume change to compensate for a density change of 1 σT unit?

Answer: Recall that 1 σT unit = 1kg/m3, or 0.001g/cm3. Let ΔV = V2 - V1, and let M and V1 be as given in Example 1. We are given that Δρ =

0.001g/cm3. Some algebra leads us to the following,

ΔV =ΔρV1

2

( M - ΔρV1 )

Using the values of M and V1 from Example 1, we compute ΔV = 49.5cm3. This result leads to the following.

Rule of Thumb: In Seaglider it takes about 50cm3 of volume change to compensate for 1 σT (1kg/m3) of density change.

Seaglider uses a system called the Variable Buoyancy Device (VBD), to change its buoyancy. This system acts as described above tochange the Seaglider's displacement without changing its mass. The Seaglider has about 800cm3 of volume change available from its VBD.Here is an example of how that available VBD range is partitioned on a typical mission.

cc

Total VBD available 800

Positive buoyancy required to expose the antenna (150)

VBD remaining 650

Negative thrust in densest water (250)

VBD remaining 400

Equivalent σT units of stratification 8

In this example, the Seaglider could fly normally with -250cc of negative displacement in the densest water, fully expose the GPS/Iridium(GPSI) antenna, and accommodate 8 σT units of density difference between the densest water and lightest water.

An important and unique feature of Seaglider is the compressibility of its pressure hull. The pressure hull is designed to have the samecompressibility as seawater, and is called an isopycnal hull. Most other oceanographic equipment is contained in pressure hulls that areessentially right circular cylinders, designed for strength and to maintain a fixed volume at all rated pressures. These pressure hulls acquirepositive buoyancy as they are pressurized, which requires compensation (subtraction of displaced volume) to maintain a prescribedbuoyancy. That same compensation has to be recovered by pumping to achieve positive buoyancy. Seaglider's isopycnal hull avoids thatneed, as the pressure hull does not acquire positive buoyancy from the compression of the surrounding seawater. For dives to 1000m, thisresults in about a 10% energy savings in the overall 24V energy budget.

3.2.2 Dynamic Forces

3.2.2.1 Lift


4 of 18

Seaglider gets lift from its body and its wings, which convert the vertical force provided by the buoyancy engine (VBD) into forward motion.Some additional lift comes from the vertical stabilizer while banked (executing turns).

3.2.2.2 DragThe Seaglider hull form is a shape that was designed to maintain laminar flow over about 70% of its hull length [Humphreys, Smith, et al.,2003]. Drag is partitioned into two types in the Seaglider flight model: induced drag and everything else (skin friction, form drag, etc.).

3.2.2.3 Hydrodynamic ModelA hydrodynamic model for Seaglider was created and verified during the development program. It is used by pilots to help with buoyancy trim,and is the mechanism by which depth-averaged currents are inferred for a Seaglider dive. The model has three parameters: lift, drag, andinduced drag, traditionally called a, b, and c. For our purposes, it is convenient to think of the hydrodynamic model as a black-box thatproduces estimates of the Seaglider's velocity as a function of computed buoyancy and observed pitch,

vmodel = F(buoyancycomputed, pitchobserved ) .

The vmodel can be resolved into horizontal and vertical components. In particular, the horizontal component, umodel, can be used with theobserved compass headings throughout a dive to determine a dead-reckoned glider track through the water. This results in a predictedsurfacing position, based on the GPS-determined dive starting point. The difference between this predicted surfacing position and the actualGPS-determined surfacing position is what provides the estimate for depth-averaged current. Similarly, the vertical component, wmodel, canbe compared with wobserved= dpdt, to adjust the VBD trim, and then to estimate vertical velocities in the water column. For details on theSeaglider hydrodynamic model, consult [Eriksen, Osse, et al., 2001].

3.2.3 Environmental

3.2.3.1 StratificationStratification is the term used to described the static stability of the ocean, with denser water below lighter water. Strong stratification means alarge change in density between two depths; weak stratification is a small change in density between two depths.

Example 3.3: Off the Washington coast in winter, the surface water has density ρ = 1.0245, while the water at 1000m depth has density ρ =1.0275g/cm3. This is a difference of 3σT units, which by our rule of thumb of section 3.2.1.2, implies that 150cm3 of volume change will berequired to maintain constant buoyancy forcing through the stratification from 1000m to the surface.

Example 3.4: Port Susan testing. Port Susan is an arm of Puget Sound, between Camano Island and the mainland. The south fork of theStillaguamish River empties into the head (northern end) of Port Susan, providing large seasonally varying freshwater inputs. Port Susan istidally mixed, and is also exposed to the southeast and northwest, so feels the prevailing winter and summer breezes. It is the primary site forSeaglider sea trials, because it has relatively low boat traffic, is 100m deep, and has a convenient launch ramp. However, it is a challengingsite in which to fly Seagliders, because of the strong tidal currents and the large stratification due to the freshwater river input. Sometimes thestratification in Port Susan is right at, or slightly beyond, the Seaglider's ability to compensate, as this example shows. Here is anotherexample of the partition of available VBD (displacement), assuming a Seaglider with 800cc of buoyancy available.

cc

Total VBD available 800

Positive buoyancy required to expose the antenna (150)

VBD remaining 650

Negative thrust in densest water (σT = 23.0) (125)

VBD remaining 525

Equivalent σT units of stratification 10.5

Minimum surface σT for normal operations 12.5 = 23.0 - 10.5

Note that the choice of maximum negative thrust in the densest water is the only pilot-changeable value in this table. And it has a practicallower limit, which is the ability to fly at the bottom of the dive. In this scenario, if the surface density is σT = 11.0, it is likely that only about halfthe antenna mast will be exposed at the surface, depending on the depth of the freshwater lens.

3.2.3.2 CurrentsDepth-averaged current over the course of a Seaglider dive influences the distance covered over the ground by Seaglider. The depth-averaged aspect of the current is important - the Seaglider can make progress towards a waypoint even in the presence of strong adversesurface currents by diving through deeper waters with more favorable currents. The maximum deep-water depth-averaged current thatSeaglider can stem is 40cm/s, or 0.8kts. That is the practical limit and requires driving the Seaglider as hard as it can be driven (in thebuoyancy sense). These dives tend to be done with large negative thrust on the dive (-350cc), and vertical velocities of about 18cm/s. Thedives take about three hours between surfacings, or about eight dives per day. The VBD pumps from stop-to-stop on each dive. This regimeuses energy at about ten times the rate of a typical dive off the Washington coast, where the dives are eight hours long (three dives per day),and the VBD stays within about half its full range. But it has been shown in two deployments in the Kuroshio that Seaglider can makecrossings of a strong western-boundary current. This is done in a triangular track, with an inshore and then an offshore transect of the strongcurrent, and a return upstream in the calmer water offshore of the strong current. One might imagine interesting tracks in the equatorialPacific that would exploit the equatorial undercurrent.


5 of 18

Vertical shear in the currents induces turning moments on the Seaglider body. Large vertical velocities (upwelling or downwelling) canintroduce large control changes in buoyancy, and in some cases, cause dives to truncate or abort prematurely.

3.3 Control of Static Forces

Seaglider's flight is controlled by specification of the location or state of three systems, which effect changes to the vehicle's pitch, roll andbuoyancy. All positions are encoded by linear potentiometers, digitized by 4096-count analog-to-digital (A/D) converters. The A/D counts runfrom 0 to 4095. Physically attainable limits (also called hardware limits) for each system are determined empirically at the time of assembly. Asafety margin is added to these physical limits to arrive at a software limit, which is that position (in A/D counts) beyond which the Seaglideroperating software will not command that particular system. Associated with each system are the following.

1. A center position, which is intended to be the vehicle neutral for that system, in a particular environment.

2. A factor that converts A/D counts to physical displacement, which is based on the mechanical design.

3. A gain that relates movement of each system to the effect it has on the Seaglider.

See the Appendix at the end of this document for detailed information.

3.3.1 Pitch

Pitch is controlled by moving the 24V battery pack forward and aft along the longitudinal axis of the Seaglider. The motion is accomplished byan electric motor, geared to drive a worm-gear in such a way that 217.39 A/D counts equals 1cm of battery mass travel ($PITCH_CNV).Seagliders typically respond to movement of the battery pack in the longitudinal axis by pitching about 15-20° per centimeter of mass travel.This $PITCH_GAIN is a parameter, as it is dependent on the particular sensor suite and trim ballast installed on each Seaglider.

Here are some typical pitch ranges and values for Seaglider.

hardware limit (A/D counts) software limit (A/D counts)

full forward (down, - ) 20 45 ($PITCH_MIN)

full aft (up, + ) 3402 3377 ($PITCH_MAX)

$C_PITCH (example) 2346

Note that A/D counts are always positive, but displacement may be positive or negative, relative to a given $C_PITCH. Pitch is usuallytrimmed so as to have about 70% of the pitch travel available for pitching down (forward of $C_PITCH), and 30% available for pitching up (aftof $C_PITCH).This is to ensure a good surface position, sufficiently pitched down to fully expose the antenna.

Example 3.5: How many centimeters is the battery pack forwardof its center position, when the Seaglider is in its surface (fully pitched-down)position?

Answer: By the values in the table above, the pitch position will be 45 A/D counts when fully forward. Given a $C_PITCH of 2346 and a$PITCH_CNV of 217.39 A/D counts/cm, we have that the pitch position = (45 counts - 2346 counts)*(1 cm/217.39 counts) = -10.585cm, or10.585cm forward of the pitch center.

Example 3.6: How many A/D counts will cause 5º of vehicle pitch? Assume a $PITCH_GAIN of 16 /cm of mass travel.

Answer: The number of A/D counts that corresponds to 5º of vehicle pitch is 5º * (1 cm/16º) * 217.39 A/D counts/cm = 67.9 A/D counts ≈ 68A/D counts.

Example 3.7: Suppose the pilot made a trim bias correction of $C_PITCH to 2375. What was the presumed pitch bias (in degrees)?

Answer: The previous $C_PITCH was 2346 as given in the table above. Then the pitch bias is computed as follows, (2346 A/D counts - 2375A/D counts) * (1 cm/217.39 A/D counts) * 16º/cm = -2.13º . The Seaglider was pitched 2.13º down.

3.3.2 Roll

Roll is controlled by rotating the 24V battery pack inside the hull. The pack is axially asymmetric, and weighted on its ventral face (as normallyinstalled). An electric motor and gear train rotate the mass such that 35.37 A/D counts is equivalent to 1 degree of battery mass rotation($ROLL_CNV). Seagliders typically respond to the rotation of the battery pack by rolling about 1/2º for every 1º of battery pack rotation, whichis also dependent on the amount and distribution of trim lead.

The control strategy is to roll the 24V battery pack a specified amount (40º) in the appropriate direction when a turn is indicated, then roll backto neutral (center) when the desired heading is reached. Note that the Seaglider turns in the opposite sense from its bank angle on the dive(opposite from upright airplane control), and in the same sense as its bank angle on the climb (same as upright airplane control). Here aresome typical roll ranges and values for Seaglider. Two roll centers are used because various asymmetries in form result in different roll trimon dives and climbs.

hardware limit (A/D counts) software limit (A/D counts)


6 of 18

full roll to port (-) 0 150 ($ROLL_MIN)

full roll to starboard (+) 3983 150 ($ROLL_MAX)

$C_ROLL_DIVE (example) 2000

$C_ROLL_CLIMB (example) 2050

3.3.3 Buoyancy

Buoyancy is controlled by a mechanism called the Variable Buoyancy Device (VBD). It is a hydraulic system whose purpose is to maintain aspecified total vehicle displacement by varying the size of an oil-filled bladder external to the Seaglider pressure hull. The system pumps oilfrom an internal reservoir into the external bladder to increase displacement, and allows oil to bleed from the external bladder into the internalreservoir to decrease displacement. Linear potientiometers on either side of the internal reservoir measure the position of the reservoirsrolling diaphragm. The mean of the two values is reported as the position of the diaphragm, which can be interpreted as the amount of oil inthe internal (or external) reservoir. The geometry of the system results in 4.0767 A/D counts per cm3 of oil ($VBD_CNV). The point of neutralbuoyancy is designated $C_VBD, and is set relative to the densest water to be encountered on a mission. VBD control is calculated toachieve specific results, which depend on pilot specified quantities: Seaglider vertical velocity, distance to next waypoint, and maximum glideslope. VBD control is the gas pedal or throttle. Specific VBD control issues will be discussed in more detail later. Here are some typical VBDranges and values for Seaglider.

hardware limit (A/D counts) software limit (A/D counts) volume (cm3)

Vmax 105 205 ($VBD_MIN) 557 (wrt$C_VBD)

Vmin 3610 3560 ($VBD_MAX) -266 (wrt$C_VBD)

Range 823

$C_VBD 2476

Note that Vmax is the maximum displaced volume of the Seaglider, and Vmin is the minimum displaced volume of the Seaglider. When given in

cm3, they are with respect to (wrt) a given $C_VBD.

Example 3.8: A typical ballasting case, which demonstrates how the available range in VBD is utilized. Values are from a recent deploymentof SG012 off the Washington coast. Assume that the $C_VBD and limit values in the table above are appropriate for a 1000m density of ρ =1.0275g/cm3. How much negative buoyancy is available at 1000m?

Answer: From the table above, -266cm3 of displacement (relative to neutral) are available at 1000m. Multiplying by the water density, thebuoyancy force available is 273g. Note that there are also 557cm3 of positive displacement available, relative to the 1000m water density. Weknow that the Seaglider antenna requires 150cm3 of positive displacement for optimal exposure. That leaves 407cm3 of displacementavailable to overcome stratification. From our rule-of-thumb, 50cm3 is equivalent to one σT unit, so 400cm3 is equivalent to 8σT units. That

means that Seaglider can get its antenna fully exposed in surface water with a density as low as (1.0275 - 0.008)g/cm3 = 1.0195g/cm3.

3.4 Features of Control

3.4.1 Canonical Dive

The Seaglider performs its mission by repeating a canonical dive until either commanded to stop or until an abort condition is reached.Numerous aspects of the canonical dive are under the control of the pilot, through an extensive set of parameters. A few are indicated in thesketch below. Many more are not shown; consult the Parameter Reference Manual for more information. The canonical dive is shownschematically in Figure 1.


7 of 18

Figure 1. A schematic of a Seaglider autonomous run profile, with run phases indicated by the intervals at the top of the figure and the profiledata boundaries indicated by the interval at the bottom of the figure. The figure is not to scale in either dimension.

3.4.2 Control Design

The Seaglider flight control scheme has two guiding principles: maintain constant vertical velocity and minimize the total energy expenditureduring a dive.

Constant vertical velocity is desired because the Seaglider samples its sensors evenly in time. Constant vertical velocity then implies thesamples are equally spaced in depth. Sample intervals are specified by the pilot through the science file. They may vary by pilot-specifieddepth bands, but are uniform within each specified depth band.

The desired vertical velocity is not specified directly by a parameter, but is calculated from parameters that describe the target depth of a dive($D_TGT) and the time to complete a dive ($T_DIVE). $D_TGT is in meters, and $T_DIVE is in minutes, and is the time from surface-to-surface., discounting pumping time at the bottom of the dive. So, the desired vertical velocity, wd = (2 * $D_TGT * 100cm/m)/($T_DIVE *60s/min) cm/s.

Example 3.9: A typical normal value for wd is 10 cm/s. What ratio of $D_TGT : $T_DIVE yields wd = 10cm/s?

Answer: A bit of simple algebra gives ($D_TGT/$T_DIVE) = (10 cm/s * 60 s/min)/(2 * 100 cm/m) = 3 m/min. So, to achieve a wd of 10 cm/s,choose $D_TGT and $T_DIVE in a ratio of 3. Our typical first shallow test dive uses $D_TGT = 45 m and $T_DIVE = 15 min.

The buoyancy and pitch used on any individual dive is chosen by the Seaglider operating software to achieve the best results on that dive.The choices are bounded by parameters $MAX_BUOY, the maximum negative buoyancy allowed on a dive, and $GLIDE_SLOPE, themaximum glide slope allowed on the dive. The choices are also bounded by physical limits, neutral buoyancy (need some negative buoyancyto glide) and the stall angle. The software has to choose a buoyancy value between 0 (neutral) and $MAX_BUOY, and a desired pitch anglebetween the stall angle and $GLIDE_SLOPE. The choice is determined by the distance to the next waypoint. The pitch angle is chosen toachieve the desired horizontal distance: maximum if the waypoint is distant, minimum if the waypoint is close, or the exact distance, ifpossible. Once the pitch angle is chosen, the buoyancy is chosen to achieve the desired vertical velocity in the densest water (deepestdepth).

The main energy draw on Seaglider is pumping hydraulic oil from the internal reservoir to the external bladder at depth, where the pump hasto overcome the seawater pressure acting on the bladder. The pump consumes about 70% of the energy budget of Seaglider. Control duringflight is generally designed to minimize the total amount of pumping required on a dive. In particular, no bleeding is allowed on descent (dive)to maintain the desired vertical velocity. Pumping as necessary is allowed on the climb to maintain the desired vertical velocity. Pitch isessentially fixed at all phases of the operation, with the exception of slight pitch maneuvers on the climb to compensate for the changes inmass distribution and buoyancy due to pumping oil from the internal reservoir into the bladder. Details of the control scheme are given in the


8 of 18

description of the run phases below.

3.4.3 Run Phases

The phases of a Seaglider autonomous run are described below. Two of the phases, launch and recovery, are generally performed at thebeginning and end of the mission. The other phases, surface, dive, apogee and climb, are meant to be repeated sequentially, once eachprofile, until the end of the mission. The surface phase is where the GPS positions are acquired, where communication is attempted andwhere the navigation calculations for the next dive are made. Various depth, time and functional triggers exist to cause the Seaglider to movefrom one phase to the next.

Data acquisition is done in the dive, apogee and climb phases of an autonomous run. During each of these phases, the glider continuouslycollects data from the scientific instrumentation at a rate set specified in the science file. Although other actions are performed during thesephases, the data collection process is never interrupted. Another periodic action performed during the profile phases (dive, apogee and climb)is guidance and control (G&C). G&C operations occur at intervals defined in the science file. Three operations occur during G&C, ifnecessary: pitch adjustment, VBD adjustment, and roll adjustment.

When G&C operations occur, the Seaglider is said to be in active G&C mode. When G&C corrections are not being made, the Seaglider issaid to be in passive G&C mode. These definitions of active and passive modes refer to G&C operations only. They do not apply to dataacquisition intervals or activities. The Seaglider is acquiring data during all profile phases, whether in active or passive G&C mode. In passiveG&C mode, the Seaglider processor enters a low-power sleep state between data acquisition points. The Seaglider flies in the state specifiedin the previous active G&C mode.

3.4.3.1 LaunchThe launch phase begins when the field operator has initiated the Sea Launch procedure, performed a self-test if indicated, and all launchdialogue has completed. (Details of the launch procedure will be given in Chapter 5, Mission Execution.) The Seaglider is in its surfaceposition (rolled to neutral, pitched fully forward, and pumped to $SM_CC, typically maximum VBD for launch), and enters a normal surfacephase: acquires GPS1, and initiates a communication session via Iridium satellite telephone.

3.4.3.2 SurfaceThe surface phase begins at the end of climb-phase data acquisition. The following steps are performed.

1. Surface Maneuver: The surface position of the Seaglider is as follows: pitched fully forward (to the software limit), rolled toneutral ($C_ROLL_CLIMB), and pumped to VBD = $SM_CC. Note that if the Seaglider surfaces with VBD > $SM_CC, no bleedingis done to force VBD = $SM_CC. There are several ways to enter the surface maneuver. The Seaglider is in the surface position atlaunch, after normal completion of a dive (reached $D_SURF), in recovery phase, or after $T_MISSION minutes have elapsedfrom the start of the dive without achieving $D_SURF in climb phase. The first test in surface phase is to see if the Seaglider depthis less than $D_SURF. If so, the Seaglider pitches fully forward and pumps to $SM_CC. If not, the Seaglider first pumps VBD to itsmaximum value, and checks the depth again. If the depth is less than $D_SURF, the Seaglider moves the pitch mass to its fullforward position. This behavior is designed to try to get the Seaglider to the surface in the event of a $T_MISSION timeout.

2. GPS1: Once the desired surface position is attained, GPS position 1 ($GPS1) is acquired. The GPS receiver is turned on andleft on until a satisfactory position is acquired or until $T_GPS minutes have elapsed. Once an initial position is acquired, theSeaglider waits an additional $N_GPS samples for a GPS position with an HDOP < 2.0. If one occurs, acquisition stops and thatposition is accepted. If one has not occured in $N_GPS samples, the last position is accepted.

3. Communications: Wireless communication via Iridium begins following acquisition (or timeout) of $GPS1. The Seaglider powersup the Iridium phone, waits a specified time for registration with the Iridium system, then attempts a data call to the basestation. Ifthe call is successful, the Seaglider logs into the basestation as a dial-up user, and uses a modified XMODEM protocol to transferfiles: data, log and engineering files from the Seaglider to the basestation, and command, control, diagnostic and special purposefiles from the basestation to the Seaglider. Details of these files and their effects are given in Chapter 4, Command and Control. Ifall file transfers were not accomplished, the Seaglider waits $CALL_WAIT seconds and tries again. It tries up to $CALL_TRIEStimes, and if not successful, continues with the surface phase, marking files as appropriate for later transfer, and incrementing the$N_NOCOMM parameter.

4. Measure surface depth and angle: After the communications session, the Seaglider computes the average of 10 pressurereadings and then the average of 10 pitch angles to obtain a measurement of the Seaglider's surface position. These values arewritten into the log file for the next dive.

5. GPS2: After the surface pressure and pitch angle average is completed, a second GPS fix, $GPS2, is acquired. This fix is themost recent position of the glider prior to diving.

6. Navigation and flight calculations: Finally are the calculations of the parameters which determine the glider flight path of the nextprofile: buoyancy, pitch angle, and heading. These computations include the Kalman filter, if enabled, and the digital bathymetrytable lookup, if enabled. Upon completion of the calculations, the surface phase is completed and a new dive phase (and newprofile) is started.

3.4.3.3 DiveThe dive phase begins upon completion of the navigation and flight calculations that conclude the surface phase. Initially, pitch is in the fullforward position and the VBD volume is equal to the endpoint of the surface maneuver. At the start of the dive phase, a VBD adjustment(bleed) only is executed during the first G&C operation to get the Seaglider as fast and deep as possible (recall that pitch is still in themaximum forward position). When the Seaglider reaches a prescribed depth, $D_FLARE, it goes into a regular G&C operation (pitch, VBD,roll) to move to the desired pitch, VBD position and course computed for the profile.

If the glider speed is too fast on the dive section of the profile (too heavy), VBD pumping is not allowed to correct the speed error. This is an


9 of 18

energy conservation consideration. As the Seaglider descends into denser water, it will become less negatively buoyant and will slow down. Ifcorrective pumping were allowed on the dive, it is possible that additional bleeding would be required to compensate as the Seagliderreached denser water. That would then mean more pumping to eventually reach the buoyancy endpoint of the surface maneuver. Excessspeed is tolerated on the dive to help minimize total energy expenditure on the profile.

In the dive phase, the Seaglider turns to starboard by banking to port (opposite to upright aircraft flight).

3.4.3.4 ApogeeWhen the target depth is reached, the Seaglider enters the apogee phase. The apogee phase is a two-G&C-cycle procedure to smoothly flyfrom the dive phase to the climb phase without stalling. During the first G&C cycle of this phase, the glider is pitched to an intermediate angle,$APOGEE_PITCH, rolled to neutral, and the VBD is pumped to 0 cc. The course adjustment and passive G&C mode are skipped. A secondG&C cycle is then executed and the glider is first pitched, then VBD is pumped, to the inverse positions of the dive (pitch = -pitch, VBD =-VBD).

Data sampling continues throughout the apogee phase.

3.4.3.5 ClimbThe climb phase begins at the completion of the second G&C cycle of the apogee phase. The Seaglider is positively buoyant and pitched up,headed for the surface at the same target vertical rate as achieved on the dive phase of the profile.

As in the dive phase, data acquisition and G&C continue at the intervals specified in the science file.

If the glider speed is too fast on the climb section of the profile (too light), VBD bleeding is NOT allowed to correct the speed error. This is anenergy conservation consideration. There are two reasons. One is that any oil that is bled will need to be pumped again during the surfacemaneuver. Second, as the glider climbs it will enter less dense water, hence becoming less positively buoyant and slowing down. VBDpumping operations are allowed in the case of the glider being too heavy and slowing down. The $MAX_BUOY restriction does not apply tothe climb phase. This will not affect the amount of energy used during the profile since the oil will be pumped anyway during the surfacemaneuver.

In the climb phase, the Seaglider turns to starboard by banking to starboard (as in aircraft flight).

When the Seaglider reaches the depth $D_SURF, it enters the passive G&C mode and continues to acquire scientific data at the specifiedinterval for a time equal to $D_SURF/(wobserved m/s), or for a maximum of 50 data points, whichever is shorter. After this period of dataacquisition, the Seaglider enters the surface phase.

3.4.3.6 RecoveryThe recovery phase is entered either by command of the pilot (when it is necessary or desireable to keep the Seaglider at the surface) or byan error condition detected by the Seaglider operating software. In the recovery phase, the Seaglider stays on the surface and acquires aseries of GPS fixes which are sent to the basestation so that the Seaglider can be recovered.

In recovery, the Seaglider enters a loop of obtaining a GPS fix and communicating every $T_RSLEEP minutes. In practice, there are abouttwo minutes of overhead in this process, so that the actual time between phone calls is closer to $T_RSLEEP + 2 minutes. This recovery loopmay be exited by sending a $RESUME directive in the cmdfile. The Seaglider will then continue diving.

Chapter 4Command and Control

4.1 Seaglider

4.1.1 Serial Communications

Direct communications with Seaglider is done by connecting one end of the supplied serial communications cable to the serial port on theSeaglider, and the other to a serial port on a computer. Any reasonable terminal emulation program should be able to talk to the glider. Portsettings are 9600 baud, 8N1, and no hardware handshaking.

4.1.2 Power on/off

The Seaglider is powered on and off by use of an external magnet, usually mounted at the end of a wand. The magnet is held in theappropriate location (marked on the forward fairing) for at least 1/2 second. The starboard side turns the Seaglider on, the port side turns theSeaglider off. (The mnemonic is "Right ON! "). Once the Seaglider is turned on, and the serial communication is established properly, a shortstart-up banner should appear on the terminal, along with a request to enter a carriage return within one minute.

4.1.3 TOM8, PicoDOS®, Seaglider Operating Program main.run

There are three "levels" of software operating on the computing hardware stack on Seaglider. The lowest level is the TOM8 monitor that runson the TT8 processor. Running above that is the PicoDOS® operating system that controls the CF8 interface between the TT8 and theCompact Flash (CF). Finally, the Seaglider operating code, nominally called main.run, runs on top of PicoDOS®, and resides on the CF.


10 of 18

Starting at the TOM8 monitor, one can start PicoDOS® by typing go 2bcf8 at the TOM8 prompt. That starts PicoDOS®. From PicoDOS®, thecommand main starts the Seaglider operating code, and enters the start-up dialog.

In a fully commissioned Seaglider, a short program is loaded into the TT8 flash that automatically executes main.run at power on, so theoperator should only see the Seaglider operating code startup dialog.

4.1.4 Menus

Interaction with the Seaglider while directly connected is by a text-based multi-level menu system. Menu items are specified either by numberor by keyword match. Item numbers and keywords are shown for each entry in the menu system. For example,

1 [param ] Parameters and configuration

is the first line of the main menu, which means that the Parameter and configuration menu can be accessed by entering 1 or param. Themenu system is simple to navigate, and multi-level jumps are permitted.

4.1.5 Extended PicoDOS®

Commands available are described in the Seaglider Extended PicoDOS® Reference Manual.

4.1.6 Files on Seaglider Compact Flash (CF)

See the Seaglider File Formats Manual for names and formats of files on the Seaglider Compact Flash.

4.2 Basestation

4.2.1 Function

The Seaglider basestation is the shoreside computer end of the Seaglider system. It has three main responsibilities. It supports a modem (ormodems) and dial-up users (Seagliders). It handles one side of the modem-to-modem file transfer protocol that moves files to and from theSeaglider. And it does the data processing necessary to produce scientific and engineering data profiles, and perform simple error-detectionand notification.

4.2.2 Configuration

The Seaglider basestation runs on a Linux platform. Various distributions have been used successfully. The basestation software packageconsists of a collection of scripts, a patched version of the XMODEM send and receive programs, and configuration instructions. Thebasestation is configured to auto-answer dial-up calls, possibly screened by callerID. Seagliders log in as normal dial-up users, and send andreceive files from their home directory. Seaglider pilots need access and write permission in those home directories in order to modifycommand and control files. Scripts are executed at login and logout that control and log various aspects of the basestation transactions.

4.2.3 Files

The Seaglider pilot interacts with four files on the basestation to command and control Seaglider: cmdfile, targets, science, andpdoscmds.bat. The basestation utilities cmdedit, targedit, and sciedit are used to modify cmdfile, targets, and science. The pdoscmds.batfile is created and modified by any text editor. See the Seaglider File Formats manual for names and formats of files on the basestation.

Chapter 5Mission Execution

5.1 Planning

Mission planning is an important part of Seaglider piloting. A basic understanding of Seaglider's operating envelope, and its strengths andweaknesses is critical to planning effective science missions. The general idea is to go far by going slow - it's the square-law dependence ofdrag on velocity that gets you. "Half a knot on half a watt" is the Seaglider motto. The sections below give the operating limits of the Seaglider.

5.1.1 Environment

5.1.1.1 StratificationAs discussed in 3.2.3.1, the range of stratification in which a Seaglider can operate normally is constrained by the total amount of VBD


11 of 18

change available, and the amount of (negative) buoyancy required for the flight plan. The pilot (or scientist) should determine the likely rangeof densities to be encountered on a proposed mission, and see if there is sufficient VBD change available to accommodate that range.Compromises can be made by reducing maximum operating depth, at the expense of duration, or by reducing thrust at apogee, at theexpense of horizontal speed.

5.1.1.2 CurrentsThe maximum depth-averaged current that Seaglider can stem is 40 cm/s, or 0.8 knots. That performance requires ballasting for 350cc ofnegative displacement, specifying vertical velocities of almost 20 cm/s, and diving to 1000m. Dives last about three hours in that case, andtotal mission length is of order six weeks. Remember that it's the depth-averaged current that counts. Surface currents can be a problem,especially when doing shallow dives (see below). Plans for crossing strong currents, such as the Kuroshio or Gulf Stream, should be carefullyconsidered, and contain both return (upstream) plans and bail-out plans.

5.1.1.3 BathymetryThe Seaglider is least efficient operating in shallow water and most efficient in deep (up to 1000m) water. The practical shallow water limit isabout 75m. It is hard to make progress toward a waypoint in water shallower than that, for three main reasons: turn radius, pump time, andsurface time. Seaglider's turning radius (a few tens of meters at typical 25 cm/s horizontal speeds) is such that a significant portion of ashallow-water dive can be spent turning onto the desired course. Seaglider's pump is optimized for efficiency at pressures equivalent to 1000m ocean depth, its rate at shallow-water pressures (about 2cc/s) means that a significant portion of a shallow-water dive can be spentpumping. And finally, the time on the surface can be a significant percentage of the dive time, and if surface currents are adverse, theSeaglider can easily lose as much distance toward a waypoint while on the surface as it gains on the dive. The UW's operating guidelines areto operate deeper than 200m on offshore (deepwater) missions, and to try to stay deeper than 75m on coastal or estuarine missions.

The current model of Seaglider is rated to 1000m depth. Deep-water target depth ($D_TGT) is typically 990m to allow for the apogeemaneuver and the discreteness of the sampling times.

Knowledge of the bathymetry of the operating area is important, too. Seaglider can read a digitized bathymetry map to determine how deep todive, or can rely on the on-board altimeter to find the bottom and initiate the apogee maneuver appropriately.

5.1.2 Science requirements

5.1.2.1 TrackSeaglider can cover about 20 km/day (12 nm) through the water in normal flight. It can station-keep within about a factor of two of the divedepth (2 km on 1 km dives, 200 m on 100 m dives). The navigation system on Seaglider is waypoint-based, not track-based. The systemdecides on the most efficient way to reach the next waypoint, but does not attempt to stay on a given track. Track-based navigation can beapproximated by using more waypoints along a desired track.

5.1.2.2 SamplingSensor sampling intervals are specified in the science file. The practical lower limit on sampling is 4 seconds. If only the conductivity andtemperature sensors are sampled, it may be possible to sample every 2 seconds, but with the oxygen and BB2F optical sensors, 4 seconds isthe lower limit. The science file also gives the ability to turn off sensors, or only energize them every nth sample, in a given depth range (orranges).

5.1.3 Endurance

Total endurance, by any metric, is dependent on many factors, including depth of dive, specified vertical velocity, type of track, stratification,pump efficiency, and communications performance. (The 24V battery pack is the limiting factor: it serves the pump and the modem.)Seaglider has completed open-ocean missions of seven months duration, in conditions of minimal stratification (Labrador Sea in winter)where power conservation was the guiding factor. Missions north of Oahu around the HOT site typically lasted four months, due tostratification, and the science requirement to resolve tides. Seaglider missions in Puget Sound are typically planned for six weeks, or less.

5.2 Preparation

5.2.1 Refurbishment

This evolution is done in the laboratory between deployments. It requires a complete opening and disassembly of the Seaglider to the majorsub-assembly level, in order to replace batteries, inspect hydraulics, motors, gear assemblies, change compact flash or SIM cards asnecessary, and other operations as necessary. Preparations for deployment following refurbishment are basically the same as those for anew Seaglider.

5.2.2 Calibrations

Compass calibrations are required each time battery packs are replaced. Sensor calibrations should be performed before and after eachdeployment, as schedule and budget permit.

5.2.3 Bathymetry maps


12 of 18

The Seaglider has the ability to read digital bathymetry maps of an operating area. These maps should be prepared and loaded on theSeaglider's compact flash prior to deployment. It is possible to upload new map files to the glider while underway, but it is easier to do itbeforehand.

5.2.4 Ballasting

This is the procedure for adding or subtracting lead weights to trim the Seaglider for the desired flight characteristics in a given densityenvironment. The initial diplaced volume determination is done in the saltwater test tank at the UW School of Oceanography. Furtherballasting for particular environments can be done using the Seaglider trim spreadsheet and a simple algebraic calculation.

5.2.5 Hardware checkout and self-test

The hardware checkout procedure is done in the lab, perhaps prior to final close-up, to ensure the Seaglider is functional. The self-test isdone from within the Seaglider operating program, and tests all Seaglider systems, including the GPS and modem. The interactive self-test isthe best way to ensure the Seaglider is working properly before proceeding.

5.2.6 Deck dives

These are simulated dives, so named because they have traditionally been done outside on a deck, so that the antenna has a clear view ofthe sky. Simulated pressure and pitch observations are generated to allow test dives ($SIM_W and $SIM_PITCH). This is a valuable way totest the end-to-end data path, since the basestation is involved and has to deal with "real" data files. The "Test Launch! " menu item (in theLaunch menu) should be used for deck dives.

5.2.7 Transport

Before packing the Seaglider for transport, use the menu option hw/misc/travel to prepare the Seaglider for transport. For short trips, theSeaglider can be transported in its wooden handling cradle. Longer trips or commercial shipment require disassembly (to the pupa level), anduse of the plastic shipping cases.

5.3 Launch

The actual Seaglider launch procedure is initiated by an operator directly connected to the Seaglider via the serial communications cable. Theoperator selects the "Sea Launch! " menu option from the Launch menu. A self-test should be executed at this time.

The pilot's responsibilities are to review the results of the self-test file, and to ensure that the proper parameters and cmdfile directives are inplace when the Seaglider is launched. The Seaglider will transfer the results of the self-test and a complete dump of all the parameters priorto launch. The pilot should carefully review these files prior to giving a positive answer when the field operator asks for clearance to launch.

Once the pilot has given clearance to launch, the Seaglider goes into a normal surface procedure: acquires GPS1 and initiates acommunications session. At that time, the usual file uploads are done, including cmdfile, and science, targets, and pdoscmds.bat, if theyexist. The normal launch procedure is to place a $QUIT directive in the cmdfile, which will hold the Seaglider on the surface while surfaceposition and acoustic ranging performance are verified. If the physical launch goes well, and the field team reports satisfactory surfaceposition and acoustic ranging, the pilot can ensure that the proper shallow (test) dive parameters are in place, then put the $RESUMEdirective in the cmdfile. That will start the first dive. Once the cmdfile has been successfully transferred, and the communications session issuccessfully completed, the $QUIT directive should be placed back in the cmdfile, so the Seaglider will hold at the surface following its firstdive.

5.4 Test and Trim Dives

5.4.1 Diveplot

Data from the initial test dive should be examined in detail prior to making any parameter changes or letting the Seaglider continue on itssecond dive. The pilot is responsible for ensuring that the Seaglider is operating properly in all respects, that trim issues (if any) are within theadjustment range of the Seaglider, and that the Seaglider is cleared to continue on its mission. Affirmative answers to these questions willallow the pilot to release the field team for other duties.

It is most efficient to plot the data from the initial dive and inspect the plots. One example of such a series of plots are those produced by aMatlab script diveplot.m.

5.4.2 Initial Revision of Trim

Initial adjustments of trim are done in order pitch, VBD, and roll. These adjustments are made by changing the system neutrals ($C_PITCH,etc.), based on regressions of controlled (or expected) quantites versus observed outcomes.

For pitch, a linear regression is made between observed vehicle pitch, as measured by the inclinometer, and pitch control, the position of thepitch mass in cm relative to $C_PITCH. This linear regression will give a pitch bias and a pitch gain. We choose to adjust the $C_PITCH sothat the pitch bias is zero. This can be done using the approach given in example 3.7. We generally only adjust half-way to the calculated


13 of 18

$C_PITCH after one dive, however, since the regressions are not always robust enough to be fully trusted after only one dive. It may takeanother short dive to get $C_PITCH reasonably close to neutral pitch.

Once pitch is close, deeper dives can be undertaken, which then will provide more steady-state flight observations. These can be used toadjust $C_VBD, if necessary. These adjustments are initially made by comparing the observed Seaglider vertical velocity, , with the Seagliderhydrodynamic model predictions of vertical velocity as a function of calculated buoyancy and observed pitch angle. The $C_VBD is adjustedto make the predicted and observed vertical velocities more closely agree. Once again, based on experience, adjustments based onregressions of small numbers of dives are done in steps of about half the predicted adjustment.

Roll trim is an ongoing process, and continues throughout the deployment. Turn rate is regressed against roll control; the idea is to trim theSeaglider to fly straight, not to trim it to be flat. Separate roll centers are used for the dive and the climb phases. But environmental conditionscan cause the Seaglider to require a lot of turning to stay on course, so the corrrect value of the roll neutral points can be elusive.

5.4.3 Progression of Dives

Typical launches progress as follows: two dives to 45m, then one or two dives to 200m-250m, then one or two dives to 500m, then full depthdives. The exact sequence depends on Seaglider performance, water depth, and other operational considerations, but the idea is to get aprogression of longer and longer dives, for better tuning of the Seaglider pitch, VBD, and roll.

5.5 Monitoring

Once the initial Seaglider trim is established on a deployment, the pilot mainly monitors the Seaglider performance. This includes how theSeaglider itself is working, and how well the Seaglider is accomplishing its scientific mission. On routine deployments, checking the Seagliderstatus once per day may be sufficient. There may be times when more frequent checking is appropriate.

5.5.1 Seaglider Performance

The basics are simple. Here are some of the things to monitor.

• Is the Seaglider communicating as expected? Check the Seaglider's comm.log file on the basestation for the latestcommunication. Do the throughput rates and communication session information seem normal (>200Bps)? Are at least two-thirdsof the communications sessions completed successfully with one phone call? Are there many missing or incomplete pieces of dataor log files?

• Are the GPS positions current? Are the GPS positions current in time and reasonable in position? Are the positions acquired inreasonable amounts of time (~45s for GPS1, ~15s for GPS2)? Are the HDOP values generally less than 2.0?

• Are the sensors functional and providing believable profiles?

• Are the Seaglider control systems (pitch, roll and VBD) working normally? Check the motor currents and rates of movement of thedevice. Is there a trend? Are there retries or errors? Regress VBD in groups of ten dives or so sequentially in time and look fortrends. Small rate hydraulic leaks have been found in this way. They manifest themselves as the Seaglider apparently gettingheavier. In fact, the Seaglider is actually losing buoyancy for a given amount of oil in the internal reservoir - oil is going somewhereother than into the external bladder.

• Is the specified surface buoyancy sufficient to fully expose the antenna? Check $_SM_DEPTHo in the p*.log file. Values of 0.7mare about normal. Anything over 1.0m should be cause to adjust $SM_CC.

5.5.2 Data Completeness and Quality

Be sure that all the data for each profile has been transferred successfully to the basestation. There are many ways to do this. One simpleway is to compare the size of the processed p*.dat file to the file size specified as the first argument in the $DATA_FILE_SIZE line in thep*.log file. If there are missing pieces, the comm.log file can be an easy way to determine which chunk of data was incompletely transferred.The resend_dive command in the extended PicoDOS® command set allows for individual chunks of data or log files to be resent at the nextSeaglider surfacing.

5.5.3 Troubleshooting

The contents of the p*.log files, the capture file mechanism, and the extended PicoDOS® command set (used via the pdoscmds.bat file)provide useful tools for troubleshooting.

5.5.4 Resources

Call the UW Seaglider Fabrication Center for further help and information.

5.6 Recovery


14 of 18

A normal Seaglider recovery is initiated by directing the Seaglider to a specified recovery point. The Seaglider pilot will often have had theSeaglider performing relatively shallow (short) dives at the recovery point while awaiting word on the arrival of the recovery team. That way,the Seaglider is relatively close to the surface at all times, and there are regular opportunities to place the Seaglider at the surface. The actualrecovery is initiated by putting the $QUIT directive in the cmdfile. An adjustment to $T_RSLEEP, the time (min) to sleep between calls whilein recovery, is often made at this time. The .pagers file mechanism is generally used to transmit the most recent Seaglider surface position tothe recovery team via SMS to an Iridium handset or cell phone.

It is important for the pilot to check the $INTERNAL_PRESSURE log file parameter prior to authorizing recovery, to check for possiblebattery outgassing.

Recoveries have also been accomplished with general locations from Iridium and specific location via acoustic ranging., when GPS was notavailable.

Physical recovery depends on the vessel in use. From low-freeboard vessels, recovery can be directly into the wooden handling cradle. Fromlarger vessels, a lasso is generally placed around the vertical stabilizer and the Seaglider is lifted aboard.

Once on deck, the Seaglider is powered off, rinsed with fresh water if available, the antenna is removed, and covers are placed on thesensors. If a computer and communications cable are available, the Seaglider can be powered on and put in the 'travel' mode as describedabove.

Chapter 6Special Topics

6.1 Scenario Run

A scenario run is a finite sequence of prescribed values for pitch, roll, and VBD. It is used primarily in tank tests to determine the displacedvolume of a Seaglider, do systems checks, or validate vehicle trim configurations. The following sections will describe the setup, phases ofoperation, and the data download for a Seaglider scenario run.

6.1.1 Setup

1. Select the "Pre-launch" menu item.

2. Select "Set scenario mode".

3. Specify the scenes by entering the desired values for pitch, roll, VBD, and duration.

4. Adjust the parameter values shown on the menu page if needed. These are the key values that have an effect in scenario runs.

5. Store the scenes if desired.

6. Return to previous menu

7. Go to the "Pre-Launch" menu again.

8. Select "Test Launch! "

6.1.2 Run Phases

A scenario run is divided into four phases: launch, scene, transition and recovery.

6.1.2.1 LaunchThe test launch dialog will get the Seaglider to the initial state, and start data collection. Once the first state (scene) is established, Once thevalues are achieved, the Seaglider starts the first scene. The mission timer starts at the completion of this countdown. It is reasonable to stayconnected to the Seaglider until the first few data acquisition cycles have been completed. Then disconnect the communcations cable, installthe connector plug and physically launch the Seaglider.

The launch phase is a special case of the transition phase between scenes, as it is a transition from the (possibly random) deck setup initialconditions to the static values specified for pitch, roll and VBD in the initial scene.

6.1.2.2 SceneA scene is a set of static values for pitch shifter position, roll shifter position, VBD volume, and duration specified during the setup of thescenario run. The Seaglider holds those values for the duration specified for the scene, then goes to the transition phase to change thosevalues to those specified for the next scene.


15 of 18

The time entered for each scene applies to the static (passive) portion of the scene only. The time in the transition phase (needed forperforming the active G&C to achieve desired pitch, roll and buoyancy) is additional.

For example, suppose we have specified a scenario with two scenes of the following pitch, roll, VBD, and duration.

Scene Pitch(cm) Roll(deg) VBD(cc) Time(sec)

1 -1.0 0 -100 60

2 1.0 0 100 60

Then the total run time will be as follows.

Phase Time

Launch Varies, depending on initial conditions of pitch, roll, and VBD

Scene 1 60 seconds

Transition (1-2) About 4 seconds for pitch and about 90 seconds to pump 200 cc

Scene 2 60 seconds

Recovery Varies, depending on time for surface maneuver

Total 4-10 minutes

6.1.2.3 TransitionDuring the transition (active) phase, the Seaglider changes the pitch position, roll position and VBD volume to the values specified for the nextscene. Once the values are attained, the next scene begins and the values are held constant for the time specified in the duration parameter.The Seaglider takes some time to stabilize following the transition phase. This is particularly true for transitions involving large pitch changes,which impart significant momentum to the Seaglider. Scene durations of at least 60 seconds are appropriate for tank testing.

6.1.2.4 RecoveryDuring a scenario run, the Seaglider goes into recovery phase if any of the following conditions are met.

1. The target depth, $D_TGT, is exceeded. Target depth in a scenario run is used like the abort depth parameter in an autonomous run.2. The abort depth, $D_ABORT, is exceeded. This should never happen because the target depth should be reached first ($D_TGT <$D_ABORT).3. The maximum mission time, $T_MISSION, is exceeded.4. All defined scenes are completed.

6.1.3 Data Access

After the run, the data can be viewed using the sumasc command from the picoDOS>> prompt, or uploaded to the lab computer using thexmodem commands at the picoDOS>> prompt.

Files could be moved to a Seaglider basestation and then converted to the normal p*.[log, asc, eng] format. If a scenario run is aborted forsome reason, the data up to the point of power-off are in the files thisdive.dat and thisdive.log.

6.2 Ballasting

A separate procedure for ballasting is in preparation.

Chapter 7Nomenclature

7.1 Terminology

Run Phase Mode Action

Scenario Launch Active (G&C) Pitch (Fwd/Aft)


16 of 18

Autonomous Surface Passive (G&C) VBD (Pump/Bleed)

Dive Sleep (TT8) Roll (Port/Stbd)

Apogee

Climb

Recovery

7.2 Conventions

Variable Buoyancy Device

Internal Bladder Condition Empty Full

Operation Pump (make more buoyant) Bleed (make less buoyant)

Overall Buoyancy Positive Negative

A/D Count < center > center

Value (cc) from Neutral Positive Negative

Pitch Mass Sensor

Mass Shifter Position Forward Aft

Operation Dive Climb


Value (cm) from Neutral Negative Positive

Roll Mass Sensor

Mass Shifter Position Port Starboard

Turn Type Starboard in dive, port in climb Port in dive, starboard in climb


Value (deg from neutral) Negative Positive

Inclinometer

Roll Port wing down Starboard wing down

Degree Reading Negative Positive

Pitch Nose down Nose up

Degree Reading Negative Positive

Appendix:Table of Conversion Factors

Pitch Roll VBD

HardwareLimits

Determined and set byassembler

Determined and set byassembler Determined and set by assembler

SoftwareLimits

$PITCH_MIN,$PITCH_MAX

$ROLL_MIN,$ROLL_MAX $VBD_MIN, $VBD_MAX

Center $C_PITCH $C_ROLL_DIVE,C_ROLL_CLIMB $C_VBD

ConversionFactors

$PITCH_CNV (217.39cts/cm)-1

$ROLL_CNV (35.37cts/deg)-1 $VBD_CNV (4.0767 cts/cc)-1


17 of 18

Gain $PITCH_GAIN ~0.5 vehicle deg/deg ~1.0 vehicle displacement/cc

Timeout $PITCH_TIMEOUT $ROLL_TIMEOUT $VBD_TIMEOUT

MaximumErrors $PITCH_MAXERRORS $ROLL_MAXERRORS $VBD_MAXERRORS

Dead Band $PITCH_DBAND $HEAD_ERRORBAND $VBD_DBAND

Rate ofMotion $ROLL_AD_RATE $PITCH_AD_RATE

$VBD_PUMP_AD_RATE_SURFACE,$VBD_PUMP_AD_RATE_APOGEE,$VBD_BLEED_AD_RATE

Copyright University of Washington, 2006


18 of 18

Documents

Seaglider Pilot's Guide